research papers
Acta Cryst. (2015). D71, 505–515 doi:10.1107/S1399004714027552 505
Received 2 September 2014
Accepted 17 December 2014
‡ These authors contributed equally to the
results and are equal co-first authors.
Keywords: arsenic(III) S-adenosylmethionine
methyltransferase; AS3MT; phenylarsenite;
roxarsone; disulfide-bond cascade.
PDB references: CmArsM, with bound
phenylarsenite, 4kw7; with bound roxarsone,
4rsr
Supporting information: this article has
supporting information at journals.iucr.org/d
A disulfide-bond cascade mechanism for arsenic(III)S-adenosylmethionine methyltransferase
Kavitha Marapakala,a‡ Charles Packianathan,b‡ A. Abdul Ajees,c Dharmendra S.
Dheeman,b Banumathi Sankaran,d Palani Kandavelue and Barry P. Rosenb*
aDepartment of Chemistry, Osmania University College for Women, Osmania University, Hyderabad 500 095, India,bDepartment of Cellular Biology and Pharmacology, Florida International University Herbert Wertheim College of
Medicine, 11200 S. W. 8th Street, Miami, FL 33199, USA, cDepartment of Atomic and Molecular Physics, Manipal
Institute of Technology, Manipal University, Manipal, Karnataka 576 104, India, dPhysical Biosciences Division,
Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94702, USA, and eSER-CAT and the Department
of Biochemistry and Molecular Biology, University of Georgia, Athens, GA 30602, USA. *Correspondence e-mail:
Methylation of the toxic metalloid arsenic is widespread in nature. Members
of every kingdom have arsenic(III) S-adenosylmethionine (SAM) methyltrans-
ferase enzymes, which are termed ArsM in microbes and AS3MT in animals,
including humans. Trivalent arsenic(III) is methylated up to three times to form
methylarsenite [MAs(III)], dimethylarsenite [DMAs(III)] and the volatile
trimethylarsine [TMAs(III)]. In microbes, arsenic methylation is a detoxification
process. In humans, MAs(III) and DMAs(III) are more toxic and carcinogenic
than either inorganic arsenate or arsenite. Here, new crystal structures are
reported of ArsM from the thermophilic eukaryotic alga Cyanidioschyzon sp.
5508 (CmArsM) with the bound aromatic arsenicals phenylarsenite [PhAs(III)]
at 1.80 A resolution and reduced roxarsone [Rox(III)] at 2.25 A resolution.
These organoarsenicals are bound to two of four conserved cysteine residues:
Cys174 and Cys224. The electron density extends the structure to include a
newly identified conserved cysteine residue, Cys44, which is disulfide-bonded to
the fourth conserved cysteine residue, Cys72. A second disulfide bond between
Cys72 and Cys174 had been observed previously in a structure with bound SAM.
The loop containing Cys44 and Cys72 shifts by nearly 6.5 A in the arsenic(III)-
bound structures compared with the SAM-bound structure, which suggests that
this movement leads to formation of the Cys72–Cys174 disulfide bond. A model
is proposed for the catalytic mechanism of arsenic(III) SAM methyltransferases
in which a disulfide-bond cascade maintains the products in the trivalent state.
1. Introduction
Arsenic is one of the most ubiquitous environmental toxins
and is a human carcinogen. It poses a serious threat to human
health, and consequently ranks first on the 2013 Environ-
mental Protection Agency’s Comprehensive Environmental
Response, Compensation and Liability Act List of Hazardous
Substances (http://www.atsdr.cdc.gov/spl/). This toxic metal-
loid is introduced primarily from geochemical sources and is
acted on biologically, creating an arsenic biogeocycle (Zhu
et al., 2014). Members of every kingdom, from bacteria to
humans, methylate arsenite, producing the trivalent species
methylarsenite [MAs(III)], dimethylarsenite [DMAs(III)] and
the volatile trimethylarsine [TMAs(III)] (Thomas & Rosen,
2013; Ye et al., 2012; Liu et al., 2013). The enzymes that
catalyze this reaction are arsenic(III) S-adenosylmethionine
(SAM) methyltransferases (AsMTs; EC 2.1.1.137), which have
been termed ArsM in microbes and AS3MT in animals. In
microbes arsenic methylation is a detoxification process (Qin
et al., 2006, 2009), but in humans the production of MAs(III)
ISSN 1399-0047
# 2015 International Union of Crystallography
and DMAs(III) has been proposed to increase reactivity and
toxicity (Thomas et al., 2007).
Whether AsMTs detoxify arsenic or transform it into more
toxic products depends on their enzymatic mechanism. Chal-
lenger proposed that the mechanism is an alternating series
of oxidative methylations and reductions, with pentavalent
species as products (Challenger, 1947). This hypothesis is
supported by the fact that humans and other mammals excrete
dimethylarsenate [DMAs(V)] and, to a lesser extent, methyl-
arsenate [MAs(V)] in urine. An alternate proposal is that the
products are all trivalent methylarsenicals (Hayakawa et al.,
2005). If the primary intracellular products of methylation
are the pentavalent species, then arsenic would have limited
carcinogenic potential. On the other hand, if the trivalent
species are the major methylated intracellular products, then
methylation would increase the carcinogenicity of arsenic.
Whether the oxidized species found in urine of mammals or
the growth medium of microbes are the products of the
AsMTs or are the result of nonenzymatic oxidation of the
unstable trivalent species is controversial (Cullen, 2014), but,
with careful handling, the trivalent forms have been detected
in urine (Le et al., 2000; Drobna et al., 2012). Thus, resolution
of this uncertainty is of considerable consequence in our
understanding of the health effects of arsenic.
In contrast to the uncertainty in the human pathway of
methylation, microbial arsenic methylation is an established
detoxification process that is proposed to have an impact on
the global arsenic cycle (Zhu et al., 2014). Genes for ArsM
orthologues of the human AS3MT are widespread in the
genomes of bacteria, archaea, fungi and lower plants (Ye et al.,
2012). ArsM from the thermoacidophilic eukaryotic red alga
Cyanidioschyzon sp. 5508 from Yellowstone National Park
(CmArsM) catalyzes arsenic methylation and volatilization,
leading to resistance (Qin et al., 2009). CmArsM is a 400-
residue thermostable enzyme (44 980 Da; accession No.
ACN39191) that methylates arsenic(III) to a final product of
volatile TMAs(III). CmArsM has been crystallized (Marapa-
kala et al., 2010) and its structure solved to 1.6 A resolution
with and without arsenic(III) or SAM (Ajees et al., 2012). To
understand on the one hand how microorganisms remodel the
environment in arsenic-rich regions and on the other hand
how arsenic methylation is involved in carcinogenesis, our goal
is to elucidate the catalytic cycle and determine the X-ray
crystal structure of CmArsM with various ligands.
Three conserved cysteine residues have previously been
recognized in ArsM orthologs, which are Cys72, Cys174 and
Cys224 in CmArsM (Marapakala et al., 2012; Supplementary
Fig. S1). Here, we report crystal structures of CmArsM with
bound phenylarsenite [PhAs(III)] and roxarsone [Rox(III)]
with final resolutions of 1.80 and 2.25 A, respectively. Both
aromatic arsenicals, which are the trivalent forms of poultry
antimicrobial growth enhancers, were found to be substrates
of CmArsM and to bind with high affinity to Cys174 and
Cys224. We now identify a fourth conserved cysteine, Cys44, in
CmArsM which, like Cys72, is required for the first methyla-
tion step but not the second. In previous structures, Cys44
was not resolved (Ajees et al., 2012). In the new structures,
Cys44 and Cys72 are found to form a disulfide bond. In the
previously reported SAM-bound structure Cys72 and Cys174
were observed to be disulfide-bonded to each other (Ajees et
al., 2012). We propose a new model for the catalytic cycle of
arsenic(III) SAM methyltransferases that utilizes a disulfide-
bond cascade to successively reduce pentavalent arsenical
intermediates. In this model the substrates and products are
trivalent, but there are pentavalent enzyme-bound inter-
mediates that are reduced by cysteine residues, creating a
disulfide-bond cascade mechanism of alternating oxidations
and reductions of the bound arsenic.
2. Materials and methods
2.1. Preparation and purification of CmArsM and derivatives
CmArsM 7B lacking 31 residues from the N-terminus and
28 residues from the C-terminus with a C-terminal histidine
tag (termed simply CmArsM) was expressed and purified as
described previously (Marapakala et al., 2010). The purified
enzyme was stored at �80�C until use. Protein concentrations
were estimated from the absorbance at 280 nm (Gill & von
Hippel, 1989).
Mutations were introduced into the CmarsM 7B gene by
site-directed mutagenesis using a QuikChange mutagenesis
kit (Stratagene, La Jolla, California, USA). The forward and
reverse primer oligonucleotides used for mutagenesis were
50-CTAAAGACGAGCGCTGCCAAGCTTGCCGCGGC
and 50-GCCGCGGCAAGCTTGGCAGCGCTCGTCTTTAG,
respectively, for Cys44 mutagenesis and 50-GTCCTGGAA-
AAGTTCTACGGTGCCGGGTCTACGC and 50-GCGTAG-
ACCCGGCACCGTAGAACTTTTCCAGGAC, respectively,
for Cys72 mutagenesis. Using these primers, the codons for the
conserved residues Cys44 and Cys72 were changed to alanine
codons, generating C44A and C72A derivatives. The double
C44A/C72A mutant was generated by mutating codon 44 in
the C72A derivative to an alanine codon using the same
forward and reverse primers. Each mutation was confirmed by
sequencing the entire gene. When the gene for each derivative
was expressed in Escherichia coli, each protein was produced
in a soluble form in amounts comparable to the parental
CmArsM. As described below, both enzymes retained the
ability to catalyze the methylation of MAs(III). These results
indicate that the substitutions did not have a major effect on
the overall folding of the proteins.
2.2. Crystallization of ArsM in the presence of aromaticarsenical ligands
CmArsM was co-crystallized with the aromatic arsenicals
PhAs(III) or Rox(III) by the hanging-drop vapor-diffusion
method as described previously (Marapakala et al., 2010;
Ajees et al., 2012). CmArsM (15 mg ml�1) was incubated with
1 mM PhAs(III) or Rox(III) for 30 min at room temperature
and 2.5 ml portions of protein solution were mixed with equal
volumes of reservoir solution consisting of 18% PEG 3350,
0.2 M calcium acetate, 0.1 M Tris–HCl pH 7.0. The hanging
drops were equilibrated with 0.5 ml well solution and crystals
research papers
506 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase Acta Cryst. (2015). D71, 505–515
were obtained at 293 K using Linbro 24-well plates (catalog
No. HR3-110; Hampton Research, Aliso Viejo, California,
USA). Crystals were transferred to a cryoprotectant solution
(25% PEG 3350, 0.2 M calcium acetate, 0.1 M Tris–HCl pH
7.0) and flash-cooled in liquid nitrogen for data collection.
2.3. X-ray data collection and processing
Data sets were collected from crystals of CmArsM with
bound PhAs(III) using an ADSC Quantum 315r (3 � 3 CCD
array) detector at 100 K under a liquid-nitrogen stream
on beamline 5.02 of the Advanced Light Source (ALS),
Lawrence Berkeley National Laboratory. Crystallographic
data for co-crystals of CmArsM with bound Rox(III) were
collected at the Southeast Regional Collaborative Access
Team (SER-CAT) facility at the Advanced Photon Source
(APS), Argonne National Laboratory. 360 image frames with
a 1� rotation angle about ’ were collected using a MAR 300
CCD detector under standard cryogenic conditions (100 K) on
synchrotron beamline 22-ID. Both data sets were indexed,
integrated and scaled using the HKL-2000 suite (Otwinowski
& Minor, 1997). Evaluation of crystal-packing parameters
indicated that the lattice can accommodate one molecule in
the asymmetric unit. The Matthews coefficients for CmArsM
with bound PhAs(III) or Rox(III) are 2.42 or 2.26 A3 Da�1,
respectively, with solvent contents of 42.0 and 45.6%,
respectively (Matthews, 1968).
2.4. Structure solution and refinement
The crystals of CmArsM with bound PhAs(III) or Rox(III)
belonged to space group C2. The unit-cell parameters for
the PhAs(III)-bound structure are a = 85.10, b = 47.41,
c = 100.34 A, � = 113.4� and those for the Rox(III)-bound
structure are a = 85.16, b = 46.41, c = 100.26 A, � = 114.1�. The
structures were determined by molecular replacement with
Phaser (McCoy, 2007). The atomic coordinates of the ligand-
free crystal structure of CmArsM (PDB entry 4fs8; Ajees et al.,
2012) were used as the starting model. Structure refinement of
each data set was performed using REFMAC5 (Murshudov et
al., 2011) as implemented in the CCP4 suite (Winn et al., 2011).
About 5% of the reflections were used for the test set. The
refined model required manual adjustment to improve the fit
to the experimental electron density using Coot (Emsley &
Cowtan, 2004). At this stage restrained refinement was carried
out and the final R factors converged to 20.5% (Rfree = 25.4%)
for the PhAs(III)-bound structure and 17.3% (Rfree = 22.4) for
the RoxAs(III)-bound structure.
A Ramachandran plot for Rox(III)-bound CmArsM
calculated using PROCHECK (Laskowski et al., 1993) indi-
cated that 91.7% of the residues are in the most favoured
region, 7.6% of the residues are in the additionally allowed
region and 0.7% of the residues are in the generously allowed
region. Final data-collection and refinement statistics and
Protein Data Bank accession codes are given in Table 1.
Figures were prepared using PyMOL (Peters et al., 2006).
LigPlot+ was used to illustrate hydrogen bonds and
research papers
Acta Cryst. (2015). D71, 505–515 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase 507
Table 1Data-collection and refinement statistics for CmArsM with boundPhAs(III) or Rox(III).
Values in parentheses are for the highest resolution bin.
PhAs(III)-bound Rox(III)-bound
Data collectionX-ray source ALS beamline 5.0.2 APS beamline 22-IDWavelength (A) 1.000 1.000Detector ADSC Quantum
315r CCDMAR 300 CCD
Resolution (A) 50.0–1.80 (1.86–1.80) 50.0–2.25 (2.29–2.25)Space group C2 C2Unit-cell parameters
(A, �)a = 85.10, b = 47.41,
c = 100.34, � = 113.4a = 85.16, b = 46.41,
c = 100.26, � = 114.1Observed reflections 135583 (11104) 120158 (4420)Unique reflections 33659 (3266) 16680 (762)Multiplicity 4.0 (3.4) 7.2 (5.8)hIi/h�(I)i 22.4 (1.7) 39.9 (6.2)Completeness (%) 98.2 (96.8) 96.8 (88.0)Rmerge† (%) 5.9 (52.5) 4.4 (24.8)Rr.i.m.‡ (%) 6.2 (17.6) 4.8 (27.0)Rp.i.m.§ (%) 3.6 (10.6) 1.8 (10.6)B factor, Wilson plot (A2) 36.9 34.8
Refinement}Rwork/Rfree†† (%) 20.5/25.4 17.3/22.4No. of protein atoms 2569 2579No. of water molecules 215 124No. of ligand atoms 9 23Mean B value (A2) 36.3 32.4R.m.s. deviation from ideal
Bond lengths (A) 0.019 0.019Bond angles (�) 1.96 1.97
Ramachandran plot‡‡ (%)Most favored 91.3 91.7Additionally allowed 8.7 7.6Generously allowed 0 0.7
PDB code 4kw7 4rsr
† Rmerge =P
hkl
Pi jIiðhklÞ � hIðhklÞij=
Phkl
Pi IiðhklÞ, where Ii(hkl) is the observed
intensity and hI(hkl)i is the average intensity over symmetry-equivalent measure-ments. ‡ Rr.i.m. =
PhklfNðhklÞ=½NðhklÞ � 1�g1=2 P
i jIiðhklÞ � hIðhklÞij=P
hkl
Pi IiðhklÞ.
§ Rp.i.m. =P
hklf1=½NðhklÞ � 1�g1=2 Pi jIiðhklÞ � hIðhklÞij=
Phkl
Pi IiðhklÞ.
} Refinement using REFMAC (Murshudov et al., 2011). †† Rwork =Phkl
��jFobsj � jFcalcj
��=P
hkl jFobsj, where Rfree is calculated for a random chosen 5% ofreflections which were not used for structure refinement and Rwork is calculated for theremaining reflections. ‡‡ Ramachandran plot calculated by PROCHECK (Laskowskiet al., 1993).
Table 2Ligand interactions with amino-acid residues.
Ligand atom Amino-acid atom Distance (A)
PhAs(III)As Cys174 O 3.1†As Cys174 C� 3.1†C6 Cys174 S 3.1†C1 Cys174 S 3.3†C2 Gly222 O 3.4†C5 Glu223 O 3.7‡C6 Glu223 O 3.8‡
Rox(III)As Cys174 C� 3.0†C4 Glu223 OE2 2.9†O3 Glu223 OE2 3.0†As Cys174 O 3.4†O1 Cys44 O 3.5‡C3 Glu223 O 3.7‡C2 Glu223 O 3.7‡C5 Gly222 O 3.7‡C6 Gly222 O 3.7‡
† Moderate hydrogen bond. ‡ Nonbonded interaction between atoms of PhAs(III) orRox(III) with amino-acid atoms.
nonbonded interactions of CmArsM with Rox(III) and
PhAs(III) (Laskowski & Swindells, 2011) and these are listed
in Table 2.
2.5. Assay of CmArsM activity
The methylation activity of purified CmArsM was assayed
at 60�C in 50 mM MOPS pH 7.5, 0.5 M NaCl containing 5 mM
GSH and 1 mM SAM as described previously (Marapakala et
al., 2012). Unless otherwise noted, the reactions were termi-
nated by adding 10%(v/v) H2O2 (final concentration) to
oxidize all arsenic species. Denatured protein was removed
by centrifugation using a 3 kDa cutoff Amicon ultrafilter. To
determine the nature of the CmArsM-bound arsenicals,
purified CmArsM (5 mM) was incubated at 60�C with 10 mM
arsenic(III) as described above for 10 min and terminated by
rapid passage through a Bio-Gel P-6 column pre-equilibrated
with the same buffer without H2O2 oxidation. Portions (25 ml)
were immediately diluted with
sufficient 6 M guanidine–HCl to
bring the final concentration to
4 M ;to release the bound
arsenicals. The products of
arsenic(III) or MAs(III) methy-
lation were separated by high-
performance liquid chromato-
graphy (HPLC) using a Hamilton
PRP-X100 C18 reverse-phase
column (Hamilton Co., Reno,
Nevada, USA) and quantified
using a PerkinElmer ELAN
DRC-e inductively coupled
plasma mass spectrometer (ICP-
MS) as described previously
(Marapakala et al., 2012). The
products of PhAs(III) methyla-
tion were separated by HPLC
using a Inertsil C4 reverse-phase
column (GL Sciences, Rolling
Hills Estates, California, USA)
with a mobile phase consisting
of 80% water at pH 1.5, 15%
methanol and 5% acetonitrile at a
flow rate of 0.3 ml min�1 and with
an oven temperature at 40�C
(Kinoshita et al., 2008).
3. Results
3.1. Structure of CmArsM withbound PhAs(III)
Crystals of CmArsM with
bound PhAs(III) diffracted to
1.80 A resolution and belonged
to space group C2, with one
molecule in the asymmetric unit
and unit-cell parameters a =
85.10, b = 47.41, c = 100.34 A,
� = 113.4�. A model of the
monomeric structure contains 328
residues from Cys44 to Ser371
(Fig. 1a). The structure includes
the N-terminal domain, the SAM-
binding domain, the arsenic(III)-
binding domain and a C-terminal
research papers
508 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase Acta Cryst. (2015). D71, 505–515
Figure 1(a) Structure of CmArsM with bound PhAs(III). Cartoon diagram (colored in light orange) representationof PhAs(III)-bound CmArsM (PDB entry 4kw7). The conserved cysteine residues are shown in ball-and-stick representation and are colored green (carbon), blue (nitrogen) or yellow (sulfur). The dark bluesphere is the As atom and the light blue sphere is a Ca2+ ion in the SAM binding site. PhAs(III) is boundbetween the conserved residues Cys174 and Cys224. (b) A disulfide bond between Cys44 and Cys72. Thefour conserved cysteine residues are shown in ball-and-stick representation. The length of the disulfidebond is approximately 2.1 A in the PhAs(III)-bound structure.
domain of unknown function (Ajees et al., 2012). Previously
reported structures started at residue Val50 in ligand-free
(PDB entry 4fs8) and arsenic(III)-bound (PDB entry 4fsd)
CmArsM and Ala49 in SAM-bound CmArsM (PDB entry
4fr0). In the PhAs(III)-bound structure, N-terminal electron
density is visible starting at Cys44. Residues 44–50 and 70–74
in the final model are well defined in the 2Fo � Fc electron-
density map contoured at 1.0� (Supplementary Fig. S3a) and
residues 44–48 form a loop that was not observed in the
previous structures. The N-terminal domain (residues 50–80)
starts with two small helices and then a long mobile loop that
we propose moves from the SAM binding sites towards the
arsenic(III) binding site during the catalytic cycle as described
below.
The electron densities of the bound PhAs(III) and the
four conserved cysteine residues are shown in Fig. 2(a). The
distance between the As atom and the thiolates of Cys174 and
Cys224 are 2.3 and 2.2 A, respectively (Fig. 1b). The distance
between the As atom and the C� atoms of Cys224 is 3.3 A. The
distance between the S atoms of Cys44 and Cys174 is 5.0 A.
The thiolates of Cys44 and Cys72 are 5.5 and 7.5 A from the
As atom, respectively.
LigPlot+ was used to identify and display the hydrogen-
bond and hydrophobic interactions of the protein with the
PhAs(III) ligand (Fig. 3a). The ligand is connected to the S
atoms of Cys174 and Cys224 and is surrounded by Cys44,
Gly222 and Glu223. The Cys44 and Gly222 residues are
involved in nonbonded interactions. The hydrogen bonds and
nonbonded interactions of bound PhAs(III) with amino-acid
residues are shown in Table 2. The CmArsM-bound PhAs(III)
structure adopts a pyramidal site in which the As atom of
PhAs(III) is coordinated to the S atoms of Cys174 and Cys224.
Each of the three liganding atom is at an
average distance of 3.3 A from the others.
Superposition of the PhAs(III)-bound
structure with ligand-free (PDB entry 4fs8,
319 C� residues), arsenic(III)-bound (PDB
entry 4fs8, 318 C� residues) and SAM-
bound (PDB entry 4fr0, 317 C� residues)
CmArsM structures yields root-mean-
square deviation (r.m.s.d.s) of 0.20, 0.41 and
1.27 A, respectively.
3.2. Structure of CmArsM with boundRox(III)
The X-ray crystal structure of CmArsM
with bound Rox(III) was determined at
2.25 A resolution. The structure belonged to
space group C2, with unit-cell parameters
a = 85.16, b = 46.41, c = 100.26 A, � = 114.1�.
CmArsM with bound Rox(III) consists of a
mixture of �-helices, �-sheets and 310-helices
similar to the arsenic(III)-bound structure
(Ajees et al., 2012). In this structure Rox(III)
binds between the conserved cysteines
Cys174 and Cys224, similar to PhAs(III)-
bound CmArsM (Supplementary Fig. S2(a).
The final model residues are well defined
in the 2Fo � Fc electron-density map
contoured at 1.0� (Supplementary Fig. S3b).
The electron densities of the bound
Rox(III) and the four conserved cysteine
residues are shown in Fig. 2(b). The
distances between the As atom and the
thiolates of Cys174 and Cys224 are 2.3 and
2.4 A, respectively (Supplementary Fig.
S2b). The other two cysteine thiolates of
Cys44 and Cys72 are 5.1 and 7.1 A, respec-
tively, from the As atom. The distance
between the S atoms of Cys44 and Cys174 is
5.0 A. The O and C� atoms of Cys174 are
approximately 3.4 and 3.0 A, respectively,
research papers
Acta Cryst. (2015). D71, 505–515 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase 509
Figure 2Stereoviews of the electron-density maps of (a) PhAs(III) and (b) Rox(III) bound toCmArsM. The stereoviews and the 2Fo � Fc electron-density maps of the conserved cysteineresidues are shown with the bound aromatic arsenicals contoured at 1.0� (light blue). Thecysteine residues, PhAs(III) and Rox(III) are shown as stick models. The As atom is coloredpurple. The PhAs(III) and Rox(III) electron densities are outlined with the Fo � Fc unbiasedOMIT map contoured at 2.5� (orange).
from the As atom. The distance between the S atoms of Cys44
and Cys174 and the As atom is 5.0 A. The Rox(III)-bound
structure adopts an overall conformation similar to that of the
arsenic(III)-bound structure. The Rox(III)-bound CmArsM
structure superimposes with the ligand-free, arsenic(III)-
bound and SAM-bound structures with r.m.s.d.s of 0.34,
0.33 and 1.32 A, respectively. This suggests that both the
PhAs(III)-bound and the Rox(III)-bound CmArsM structures
are similar to the ligand-free and arsenic(III)-bound structures
and that the SAM-bound structure adopts a different
conformation. Similar to the PhAs(III)-bound structure,
Rox(III) is sandwiched between Cys44, Glu223 and Gly222
(Fig. 3b). The hydrogen-bond and nonbonded interactions
of Rox(III) with nearby residues may serve to stabilize the
Rox(III)-bound form (Table 2). The Rox(III)-bound
CmArsM structure reveals a pyramidal site similar to that
in the PhAs(III)-bound structure, with each of the three
liganding atoms at an average distance of 3.4 A from the
others. These organoarsenical derivatives are similar to the
product of the first methylation step, MAs(III), with an
aromatic group instead of a methyl group. When arsenic(III)
accepts a methyl group from SAM, the methyl group must
undergo an inversion of configuration to allow a second
methylation. The methyl group would then be facing the bulk
solvent. In both the PhAs(III) and Rox(III) structures the
aromatic R groups are exposed to bulk solvent, indicating
that they are more similar to product-bound forms than to
substrate-bound forms.
The PhAs(III)-bound and Rox(III)-bound structures can
be superposed with an r.m.s.d. of 0.36 A over 324 aligned C�
residues (Supplementary Fig. S5a). Superposition of the
arsenic(III)-bound structure with the two structures with
bound aromatic arsenicals show r.m.s.d.s of 0.40 A over 318
aligned C� residues and 0.32 A over 322 aligned C� residues,
respectively (Supplementary Fig. S5b). The C atoms of the
aromatic rings of PhAs(III) and Rox(III) interact with resi-
dues Gly222 and Glu223 neighboring the arsenic(III)-binding
site. These interactions may reduce the access of SAM to its
binding site and may explain our inability to obtain crystals
with both arsenical and SAM bound.
3.3. The N-terminal extension and Cys44–Cys72 disulfidebond
Cysteine residues are frequently evolutionarily conserved
and play essential roles in the structure and function of
proteins such as the formation of disulfide bonds that stabilize
the structure (Fomenko et al., 2007; Beeby et al., 2005). In both
the PhAs(III)-bound and Rox(III)-bound structures a disul-
fide bond is observed between Cys44 and Cys72 with an
average distance of 2.1 A (Fig. 2). The typical S—S bond
research papers
510 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase Acta Cryst. (2015). D71, 505–515
Figure 4The MAs(III)–enzyme intermediate. Purified CmArsM (5 mM) wasincubated at 60�C with 10 mM arsenic(III) containing 5 mM GSH withor without 1 mM SAM. After 10 min, samples were separated using aBio-Gel P-6 spin column. Portions (25 ml) were diluted with 6 M Gu–HClto a final concentration of 4 M to denature the protein and release theenzyme-bound arsenic. After speciation and quantification by HPLC–ICP-MS, 0.86 mol MAs(III) was bound per mole of CmArsM.
Figure 3Residues interacting with PhAs(III) or Rox(III). LigPlot+ representa-tions illustrating hydrophobic and hydrogen-bond interactions with (a)PhAs(III) (PDB entry 4kw7) and (b) Rox(III) (PDB entry 4rsr). Ligandbonds are shown in purple and the hydrogen bond is shown in olive green.Nonligand residues involved in hydrophobic contacts are shown as arcscolored brick red. All other atoms are shown as spheres and are coloredblack (carbon), blue (nitrogen), red (oxygen), yellow (sulfur) andmagenta (arsenic).
distance in a protein structure is about 2.05 A, indicating that
the two residues are linked by a disulfide bond. Disulfide-bond
formation may restrict the movement of the N-terminal loop
observed when SAM is bound, which may also contribute to
the inability to obtain crystals with both arsenical and SAM
bound (Ajees et al., 2012). Why was Cys44 not observed in the
arsenic(III)-bound structure reported previously (Ajees et al.,
2012)? An additional density was observed near Cys72 in the
arsenic(III)-bound structure that was originally interpreted to
be a glycerol molecule (Ajees et al., 2012), but is more likely to
be a poorly resolved Cys44 in a disulfide bond with Cys72.
Why is the region more easily resolved in the PhAs(III)-bound
and Rox(III)-bound structures compared with the arsenic(III)-
bound structure? During the first round of the refinement
process of PhAs(III)-bound and Rox(III)-bound CmArsM,
the electron-density map (2Fo � Fc) contoured at 1.0� cutoff
traces the residues in the N-terminal region (residues 49–44).
In contrast, in the arsenic(III)-bound structure (Ajees et al.,
2012) the density traces of the residues are visible below a 0.5�cutoff after multiple rounds of refinement. This may reflect a
more ordered structure with the aromatic trivalent arsenicals,
which bind to the arsenic binding sites with higher affinity than
inorganic arsenic(III).
We previously reported that after 10 min
of reaction at 60�C CmArsM methylated
arsenic(III) primarily to soluble MAs(III),
but that most of the arsenic remained bound
to the enzyme at early times (Marapakala et
al., 2012). At later times nearly all of the
arsenic was transformed into soluble
DMAs(V). We considered the possibility
that MAs(III), the product of the first
methylation step, remains enzyme-bound
and is released only after the second round
of methylation. To examine the nature of the
enzyme-bound arsenic, arsenic(III) methy-
lation was carried out for 10 min at an
enzyme:substrate ratio of 1:2 to allow the
isolation of sufficient enzyme–product
complex. Approximately 43% of the added
arsenic was bound to the enzyme, which
represents 0.86 mol bound arsenic per mole
of enzyme. Unbound arsenic was removed,
and the protein with bound arsenicals was
denatured with guanidine–HCl to release
the protein-bound arsenicals (Fig. 4).
Denatured protein was removed by filtra-
tion, and the filtrate was speciated by
HPLC-ICP-MS. All of the bound arsenic
was MAs(III), consistent with our hypoth-
esis that the product of the first round of
methylation is enzyme-bound until methy-
lated a second time.
3.4. Movement of the N-terminal domain
Superposition of the SAM-bound struc-
ture with the PhAs(III)-bound and
Rox(III)-bound structures gave r.m.s.d.s of
1.27 and 1.32 A, respectively (Supplemen-
tary Fig. S5c). Comparison of these struc-
tures suggests that there is a 6.5 A
movement of the N-terminal residues 50–80
of the N-terminal domain of SAM-bound
CmArsM toward the PhAs(III)-binding or
Rox(III)-binding domain (Fig. 5a, Supple-
mentary Fig. S4a). This observed reorienta-
tion of the N-terminal loop may reflect the
research papers
Acta Cryst. (2015). D71, 505–515 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase 511
Figure 5(a) Superimposition of PhAs(III)-bound CmArsM (light orange) with the SAM-boundstructure (light grey) gives an r.m.s.d. of 1.27 A. We propose that a 6.5 A shift in the N-terminalloop from its position in the SAM-bound structure to its position in the aromatic arsenical-bound structures leads to disulfide-bond formation between Cys72 and Cys174. PhAs(III) andcysteine residues are shown in ball-and-stick representation with coloring of atoms as in Fig. 1.SAM is shown in ball-and-stick representation and colored purple (carbon), blue (nitrogen)and red (oxygen). (b) Modeling the complex of CmArsM with SAM and PhAs(III). Theternary complex of CmArsM with bound SAM and PhAs(III) was modeled by superposition oftheir individual structures. Coloring is as in Fig. 5(a). The distances from the S-methyl group ofSAM to the S atoms of conserved cysteine residues and to the As atom of PhAs(III) areindicated.
conformational changes that occur in different steps of the
catalytic cycle. The distance between the sulfur thiolate of
Cys72 when PhAs(III) is bound compared with Cys72 when
SAM is bound is approximately 5.8 A. The S atoms of Cys44,
Cys72, Cys174 and Cys224 in the PhAs(III)-bound structure
are 6.4, 7.6, 5.3 and 5.6 A, respectively, from the S-methyl
group of SAM. Similarly, the S-methyl group of the SAM is
6.1, 7.4, 5.6 and 5.8 A from the S atoms of Cys44, Cys72,
Cys174 and Cys224, respectively, of the Rox(III)-bound
structure. The S-methyl group is poised for transfer to the As
atoms of PhAs(III) or Rox(III), which are 4.0 and 4.2 A,
respectively, from the C atom of the S-methyl group (Fig. 5,
Supplementary Fig. S4b).
3.5. The role of Cys44 in the methylation activity of ArsMand the C72A/C44A double mutant
We previously examined the function of the three
conserved residues Cys72, Cys174 and Cys224 (Marapakala et
al., 2012). Mutagenesis of any of the three residues eliminated
the first round of methylation of arsenic(III) to MAs(III), but
a C72A derivative was able to carry out the second step of
methylation of MAs(III) to the dimethyl species. Cys44 is here
reported as a fourth conserved cysteine, forming a disulfide
bond with Cys72 when the arsenic-binding site is filled. A
C44A derivative was constructed, expressed and purified. The
effect of this cysteine-to-alanine substitution on the methyla-
tion of arsenic(III) and MAs(III) was examined (Fig. 6a). The
C44A protein was unable to methylate arsenic(III), which
is similar to the effect of mutagenesis of the other three
conserved cysteine residues. Like the C72A derivative, the
C44A protein retained the ability to methylate MAs(III). A
double C44A/C72A mutant gave similar results, methylating
MAs(III) but not arsenic(III) (Fig. 6b). These results suggest
that Cys44 and Cys72 play a role in catalysis that is different
from those of the arsenic-binding residues Cys174 and Cys224.
3.6. CmArsM methylates PhAs(III)
CmArsM has been shown to methylate arsenic(III),
MAs(III) and Sb(III) (Marapakala et al., 2012) and is here
shown to bind aromatic arsenicals. There are no reports of the
methylation of aromatic arsenicals by AsMTs, but a methy-
lated derivative of PhAs(III) has been observed in the urine of
patients accidentally exposed to phenyl arsenicals (Kinoshita
et al., 2008). CmArsM was incubated with PhAs(III) under
catalytic conditions, and the formation of a new species was
observed (Supplementary Fig. S6). PhAs(III), PhAs(V) and
the putative methylated species PhMAs(V) were speciated by
HPLC using a C4 column (Kinoshita et al., 2008). The new
species eluted at a position corresponding to the previously
reported position of PhMAs(V). PhMAs(V) is likely to be
the result of nonenzymatic oxidation of the trivalent product
PhMAs(III). In the absence of enzyme, PhAs(III) oxidizes to
PhAs(V). This is the equivalent of the nonenzymatic oxidation
of DMAs(III) to DMAs(V) in urine (Le et al., 2000). These
results are consistent with CmArsM methylating aromatic
arsenicals.
4. Discussion
Arsenic is considered the most ubiquitous environmental
toxin and carcinogen, and hundreds of millions of people
worldwide are exposed to arsenic in their food and drinking-
water supplies. In humans, exposure is a major contributor
to arsenic-related diseases (Abernathy et al., 2003; Tchounwou
et al., 2003), including bladder (Chen et al., 2003), lung (Putila
& Guo, 2011) and skin cancers (Rossman et al., 2001). A major
question is whether the risk is from exposure to inorganic
arsenicals or to the products of methylation by AsMTs. The
answer depends in a large part on deducing the enzymatic
mechanism of this class of enzymes. In the 1940s Challenger
proposed that the pathway of arsenic methylation involved
a series of oxidative methylations such that the substrates
research papers
512 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase Acta Cryst. (2015). D71, 505–515
Figure 6Cys44 is required for the methylation of arsenic(III) but not of MAs(III).Methylation of arsenic(III) or MAs(III) was assayed for 2 h at 60�C asdescribed in x2 with either (a) arsenic(III) or (b) MAs(III) at 10 mM.Samples were analyzed by reverse-phase HPLC–ICP-MS.
were trivalent arsenicals and the products were pentavalent
(Challenger, 1947). Hayakawa et al. (2005) proposed that both
the substrates and the products are trivalent. This difference
between the two pathways is key to whether the methylated
products are the relatively nontoxic and noncarcinogenic
pentavalent species or the more toxic and carcinogenic tri-
valent species. Cullen recently reviewed the chemistry of
methylation and concluded that ‘the Challenger pathway with
SAM as a donor of CH3+ remains the most rational option to
describe the biological methylation of arsenic’ which might
suggest a positively charged pentavalent methylarsenic
enzyme-bound intermediate (Cullen, 2014). However, no one
has demonstrated the presence of a pentavalent arsenic
intermediate bound to an arsenic(III) SAM methyl-
transferase. This question remains controversial and will
require detailed structure–function and enzymological
analysis to settle.
What can be learned from over 50 years of study of other
members of the methyltransferase superfamily? They all
append a methyl group from SAM to acceptor groups in SN2
displacement reactions involving the attack of an electron-
rich, methyl-accepting atom, usually O, N, C or S, on the SAM
methyl group, with the release of SAH and inversion of the
configuration of the methylated intermediate (Liscombe et al.,
2012). The SAM-binding fold is conserved, and they utilize a
common enzymatic mechanism. In contrast to the proposed
oxidation of As(III) in the Challenger pathway, no O-, N-, C-
or S-SAM methyltransferase catalyzes substrate oxidation.
Lysine and arginine methyltransferases, which, like AsMTs,
also methylate their substrate three times, do so without
oxidation of the nitrogen. Thus, as a class, these enzymes
methylate their substrates without substrate oxidation and, by
application of the principle of parsimony, there is no reason
to consider that AsMTs would have evolved a mechanism
research papers
Acta Cryst. (2015). D71, 505–515 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase 513
Figure 7Proposed reaction scheme for AsMTs. The proposed reaction pathway is a general one for AsMTs, but the numbering of the cysteine residues follows thesequence of CmArsM. (1) In the first round of methylation AsMT binds arsenic(III) in a series of three thiol-transfer reactions. (2) The methyl group ofSAM is attacked by the arsenic lone pair. (3) A positively charged pentavalent MAs(V) intermediate is formed and (4) reduced to an enzyme-boundMAs(III) intermediate by Cys44, with the formation of a Cys44–Cys72 disulfide. (5) The disulfide is reduced by Trx, regenerating the enzyme for thesecond round of methylation. (6) MAs(III), which is tightly bound to Cys174–Cys224, is methylated, producing an enzyme-bound positively chargedpentavalent DMAs(V) intermediate (7) which is reduced to DMAs(III) by Cys72 with the formation of a Cys72–Cys174 disulfide. (8) DMAs(III), whichis weakly bound to the single Cys224, dissociates from the enzyme and is oxidized nonenzymatically to DMAs(V). Finally, the Cys72–Cys174 disulfide isreduced by Trx, allowing the cycle to begin over.
different from other members of the superfamily. However,
unlike the nonmetal (O, N, C and S) substrates, the metalloid
arsenic is redox-active under physiological conditions, so a
positively charged enzyme-bound intermediate is a possibility,
either as an obligatory step in the catalytic cycle or as a side
reaction owing to oxidation.
The structure–function study reported here was designed to
address this issue. Based in part on the two different disulfide
bonds observed in CmArsM structures, we propose a new
and novel reaction scheme involving the formation of two
sequential disulfide bonds between Cys44 and Cys72 and then
between Cys72 and Cys174 (Fig. 7). Since we have no data
about the third methylation, the proposed pathway considers
only the first two methylations in eight steps. In the first round
of methylation (1) CmArsM binds arsenic(III) in a series of
three thiol-transfer reactions from As(GS)3, the preferred
substrate (Marapakala et al., 2012). (2) The methyl group of
SAM is attacked by the arsenic lone pair. (3) A positively
charged pentavalent MAs(V) intermediate is formed. (4)
Cys44 provides electrons to reduce the enzyme-bound
MAs(V) intermediate to MAs(III), allowing the second round
of methylation. Consistent with the postulation of an enzyme-
bound intermediate, most of the arsenic is enzyme-bound as
MAs(III) (Fig. 4) and little is released into the medium
(Marapakala et al., 2012) within the first 10 min of methyla-
tion. (5) Oxidized Cys44 forms a disulfide bond to Cys72,
which must be reduced before the next round of methylation
can occur. In vivo, Trx has been suggested to be the reductant
for AsMTs (Thomas et al., 2004), but in these studies GSH was
used as the reductant because thermostable thioredoxin and
thioredoxin reductase were not available. MAs(III) remains
strongly bound by the thiol pair Cys174–Cys224. (6) A second
methylation forms a positively charged pentavalent DMAs(V)
intermediate which is (7) reduced to DMAs(III) by Cys72,
forming a Cys72–Cys174 disulfide. (8) This disulfide is reduced
by Trx, regenerating the enzyme and releasing DMAs(III),
which nonenzymatically oxidizes in air to DMAs(V). Thus, the
four conserved cysteine residues play two distinct roles, firstly
as the binding site for arsenicals and secondly as a source of
electrons to maintain arsenic in the reduced form.
This pathway explains nearly all of the current results with
AsMTs. It resolve the differences between the Challenger
hypothesis of oxidative methylations (Challenger, 1947), but
with enzyme-bound pentavalent intermediates (Cullen, 2014),
and the idea of Hayakawa and coworkers that the substrates
and products are trivalent (Hayakawa et al., 2005). It is not
clear whether pentavalent intermediates are obligatory inter-
mediates or side products. In either case, the arsenic must be
reduced to trivalency before the reaction can proceed. The
electrons to reduce the pentavalent intermediates come from
conserved cysteine residues, which form disulfide bonds. Trx is
proposed to be involved (Thomas et al., 2004), but in this new
model it serves in the classical role of thioredoxins in disulfide-
bond reduction (Holmgren, 1989). Thus, AsMTs employ a
basic catalytic mechanism similar to that of O-, N-, C- and
S-methyltransferases. What differentiates AsMTs from other
members of the methyltransferase superfamily is the necessity
to bind trivalent arsenicals and to maintain them in the
reduced form, for which they utilize four conserved cysteine
residues.
Acknowledgements
This work was supported by NIH grant R37 GM55425. This
project utilized the Southeast Regional Collaborative Access
Team (SER-CAT) 22-ID beamline of the Advanced Photon
Source, Argonne National Laboratory. Use of the Advanced
Photon Source was supported by the US Department of
Energy, Office of Science, Office of Basic Energy Sciences
under contract No. W-31-109-Eng-38. The Berkeley Center
for Structural Biology is supported in part by the National
Institutes of Health, National Institute of General Medical
Sciences and the Howard Hughes Medical Institute. The
Advanced Light Source is supported by the Director, Office of
Science, Office of Basic Energy Sciences of the US Depart-
ment of Energy under Contract No. DE-AC02-05CH11231.
References
Abernathy, C. O., Thomas, D. J. & Calderon, R. L. (2003). J. Nutr. 133,1536S–1538S.
Ajees, A. A., Marapakala, K., Packianathan, C., Sankaran, B. &Rosen, B. P. (2012). Biochemistry, 51, 5476–5485.
Beeby, M., O’Connor, B. D., Ryttersgaard, C., Boutz, D. R., Perry,L. J. & Yeates, T. O. (2005). PLoS Biol. 3, e309.
Challenger, F. (1947). Sci. Prog. 35, 396–416.Chen, Y.-C., Su, H.-J., Guo, Y.-L. L, Hsueh, Y.-M., Smith, T. J., Ryan,
L. M., Lee, M.-S. & Christiani, D. C. (2003). Cancer Causes Control,14, 303–310.
Cullen, W. R. (2014). Chem. Res. Toxicol. 27, 457–461.Drobna, Z., Del Razo, L. M., Garcıa-Vargas, G. G., Sanchez-Pena,
L. C., Barrera-Hernandez, A., Styblo, M. & Loomis, D. (2012). J.Expo. Sci. Environ. Epidemiol. 23, 151–155.
Emsley, P. & Cowtan, K. (2004). Acta Cryst. D60, 2126–2132.Fomenko, D. E., Xing, W., Adair, B. M., Thomas, D. J. & Gladyshev,
V. N. (2007). Science, 315, 387–389.Gill, S. C. & von Hippel, P. H. (1989). Anal. Biochem. 182, 319–
326.Hayakawa, T., Kobayashi, Y., Cui, X. & Hirano, S. (2005). Arch.
Toxicol. 79, 183–191.Holmgren, A. (1989). J. Biol. Chem. 264, 13963–13966.Kinoshita, K., Noguchi, A., Ishii, K., Tamaoka, A., Ochi, T. & Kaise,
T. (2008). J. Chromatogr. B, 867, 179–188.Laskowski, R. A., MacArthur, M. W., Moss, D. S. & Thornton, J. M.
(1993). J. Appl. Cryst. 26, 283–291.Laskowski, R. A. & Swindells, M. B. (2011). J. Chem. Inf. Model. 51,
2778–2786.Le, X. C., Lu, X., Ma, M., Cullen, W. R., Aposhian, H. V. & Zheng, B.
(2000). Anal. Chem. 72, 5172–5177.Liscombe, D. K., Louie, G. V. & Noel, J. P. (2012). Nat. Prod. Rep. 29,
1238–1250.Liu, Z., Rensing, C. & Rosen, B. P. (2013). Metals in Cells, edited by V.
Culotta & R. A. Scott, pp. 429–442. Hoboken: John Wiley & Sons.Marapakala, K., Ajees, A. A., Qin, J., Sankaran, B. & Rosen, B. P.
(2010). Acta Cryst. F66, 1050–1052.Marapakala, K., Qin, J. & Rosen, B. P. (2012). Biochemistry, 51,
944–951.Matthews, B. W. (1968). J. Mol. Biol. 33, 491–497.McCoy, A. J. (2007). Acta Cryst. D63, 32–41.Murshudov, G. N., Skubak, P., Lebedev, A. A., Pannu, N. S., Steiner,
R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011).Acta Cryst. D67, 355–367.
Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307–326.
research papers
514 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase Acta Cryst. (2015). D71, 505–515
Peters, B., Moad, C., Youn, E., Buffington, K., Heiland, R. & Mooney,S. (2006). BMC Struct. Biol. 6, 4.
Putila, J. J. & Guo, N. L. (2011). PLoS One, 6, e25886.Qin, J., Lehr, C. R., Yuan, C., Le, X. C., McDermott, T. R. & Rosen,
B. P. (2009). Proc. Natl Acad. Sci. USA, 106, 5213–5217.Qin, J., Rosen, B. P., Zhang, Y., Wang, G., Franke, S. & Rensing, C.
(2006). Proc. Natl Acad. Sci. USA, 103, 2075–2080.Rossman, T. G., Uddin, A. N., Burns, F. J. & Bosland, M. C. (2001).
Toxicol. Appl. Pharmacol. 176, 64–71.Tchounwou, P. B., Patlolla, A. K. & Centeno, J. A. (2003). Toxicol.
Pathol. 31, 575–588.Thomas, D. J., Li, J., Waters, S. B., Xing, W., Adair, B. M., Drobna, Z.,
Devesa, V. & Styblo, M. (2007). Exp. Biol. Med. (Maywood), 232,3–13.
Thomas, D. J. & Rosen, B. P. (2013). Encyclopedia of Metalloproteins,edited by R. H. Kretsinger, V. N. Uversky & E. A. Permyakov, pp.138–143. New York: Springer.
Thomas, D. J., Waters, S. B. & Styblo, M. (2004). Toxicol. Appl.Pharmacol. 198, 319–326.
Winn, M. D. et al. (2011). Acta Cryst. D67, 235–242.Ye, J., Rensing, C., Rosen, B. P. & Zhu, Y.-G. (2012). Trends Plant Sci.
17, 155–162.Zhu, Y.-G., Yoshinaga, M., Zhao, F.-J. & Rosen, B. P. (2014). Annu.
Rev. Earth Planet. Sci. 42, 443–467.
research papers
Acta Cryst. (2015). D71, 505–515 Marapakala et al. � Arsenic(III) S-adenosylmethionine methyltransferase 515
Top Related